GRUAN ICM-12, Virtual meeting, Nov. 18, 2020

Takuji Sugidachi, (Meisei Electric Co. LTD., Japan)

1

Status of the Meisei SKYDEW instrument

Output data toMeisei Radiosonde

orOther radiosonde(XDATA)

Peltier

Optical sensor

PT100

2

Peltier-based chilled-mirror hygrometer “SKYDEW”

Features of SKYDEW hygrometer1. Two-stage Peltier device

No need to use cryogen material (CHF3)

2. Dew/frost detection by scattered light using an electronically modulated light

3. Digital controller (PID controller, gain scheduling depending on dewpoint)

4. Meisei original data format or XDATA format

Sensor

Heatsink

Control board

Battery

SKYDEW has been developed since 2009

by Meisei and Hokkaido university.

Optical fiber

IceMirror

size 2 – 5 um

at -60degC

3

History of SKYDEW development

2009 – 2014 Phase 1:

Several types of prototype

Lab tests and 9 test soundings

2016 – 2019 Phase 2:

Design of product model

20 test soundings including the

comparison sounding with CFH

2019 – 2020 Phase 3:

Test of 2nd product model

Product release for domestic users (in Japan)

--> These results are described in Sugidachi, 2014,

Ph. D paper at Hokkaido Univ.

(https://eprints.lib.hokudai.ac.jp/dspace/bitstream/21

15/55416/1/Takuji_Sugidachi.pdf)

We have conducted many lab tests and test sounding with some

type of prototypes.

Evaluation of the Peltier cooling performance, and development

of an additional cooling system with ethanol evaporation

Tuning the feedback (PID) controller to maintain the constant

dew/frost by trial & error, and the theoretical consideration.

9 test soundings in Japan and Indonesia.

1st prototype

Observing the dew/frost on the mirror

Cooling system with ethanol

4

History of SKYDEW development

2009 – 2014 Phase 1:

Several types of prototype

Lab tests and 9 test soundings

2016 – 2019 Phase 2:

Design of product model

20 test soundings including the

comparison sounding with CFH

2019 – 2020 Phase 3:

Test of 2nd product model

Product release for domestic users (in Japan)

--> Refer to the presentations of ICM-10/11 by Sugidachi.

20 test soundings with 1st product model

For the daytime, the measurement range is limited by the Peltier

cooling performance.

SKYDEW

CFH

SKYDEW

CFH

Nighttime

Lindenberg

1 April 2019

Daytime

Tateno

20 April 2019

Cooling limit

2nd product model

After the test sounding at Lindenberg in 2019, SKYDEW was

redesigned for better performance and productivity.

Performance: Cooling efficiency of heat sink for Peltier device.

Because of the improved Peltier cooling, SKYDEW can measure

up to 30km without additional cooling by ethanol system.

We found that the ethanol generate some harmful effects for

the stratospheric measurement. (e.g., contamination)

--> No ethanol model

Production issues solved:

Connection between a Peltier and a mirror part.

The incomplete connection leads the insufficient heat conduction,

which cause the lower cooling limitation or unstable PID control.

This issue was solved by the original specialized soldering

techniques

No ethanol

5

Specialized soldering method for the strong

connection between Peltier and mirror.

Test flight of 2nd product model

1st product model faced a challenge about the cooling limitation.

Even when the additional cooling by an ethanol evaporation is used,

the daytime performance was limited by the Peltier capability.

The sounding with 1st model @Lindenberg

Cooling limit

The hot side of Peltier

(= heat sink) become

too hot to measure the

stratospheric frost point.

Test of heatsinks

(in a chamber)No Heat sink

1st model

2nd model

Large heat sink

Hot-side

temperature

Cold-side

temperaturePeltier cooling start

(current = 2A)

When Peltier current started, the heat sink temperature

becomes gradually warmer than ambient air. The

increase rate depend on the heat sink efficiency.

The 2nd model does NOT need ethanol cooling system

6

Cooling limit

Performance of 2nd product model

7

Cooling limit

The SKYDEW measured successfully up to 29km at nighttime, and up to 25km at daytime.

For both soundings, SKYDEW use no additional cooling with ethanol.

FL31 Nighttime FL35 Daytime

* The FL35 data is provided by JMA

No ethanol cooling system

Specification

8

600g balloon

7.5m string

Parachute

30m unwinder

Typical configuration

Size & WeightDimensions 170( W) x 93 (D) x 318 (H) mm

Weight ~700 g (including Battery, without radiosonde RS-11G)

Power Source Voltage(Model)

9 VDC for Peltier device(Lithium battery (AAA)×6)

9VDC for electric circuit such as optical sensor(Lithium battery (006P)×1)

In the case of no ozonesonde

SKYDEW

Operation flow

9

4. Build up the payload, and release

1. Connect the battery (006P) to start working

for electric circuit, and connect the radiosonde

2. Check the level of scattered light signal to confirm

the mirror condition.

(When the signal is higher than the defined value,

the clean up of mirror is needed because the mirror

may be dirty)

3. Connect the battery(AAA) for Peltier to start

cooling mirror, and then mirror temperature

become the dewpoint temperature.

Battery Radiosonde

In the case with the Meisei radiosonde

For XDATA version (RS41), the decoding

software is developed by GRUAN L/C.

Battery for Peltier

Meisei software

Mirror temp.

Scattered light

Nothing just

above mirror

SKYDEW

Base + 0.3V = target

Data processing for SKYDEW toward GDP

The FL31 is a good sample for validation because this sounding obtained the descend data successfully below 22km.

By comparing the ascent and descent data, we have evaluated the performance at UTLS measurements.

Ascent Descent

Cooling limit

Cooling limit

Below 22km, both ascent and

descent data are available.

Good !Good !

10

Data processing for SKYDEW toward GDP (under consideration)

11

A raw mirror temperature measured by chilled-mirror hygrometers often has oscillations by the PID controller to

maintain a constant frost layer. These oscillations should be smoothed out by an appropriate filter (Vömel et al., 2016)

The required data processing:

1. Smoothing such as Gaussian filter

2. Time lag correction (Further evaluations are needed)

3. Others (if needed) Time lag correction

tau= 10 sec (Tentative value)No time lag correction

Raw data

Smoothed data

(Ascent)

Smoothed data

(Descent)

Uncertainty budget for SKYDEW toward GDP(under consideration)

12

(1) Calibration of temperature sensor (PT100) <0.06 K, calibration point -95, -45, 0, 40 degC

(2) Instability of PID controller < ?? K

Does a large oscillation cause the bias by unbalance of cool/heat power? (--> refer to the ICM-11 ppt)

(3) Time lag effect ?? sec

Delay by an evaporative/condensation rate of ice. Negligible for SKYDEW, depending on ice size.

Effect of slight residual water/ice attached on the sensor cover.

(4) Contamination error under cloudy condition

The clear sky condition is recommended for stratospheric measurement. (--> next slide)

(5) Others

ambiguity of the condensate phase (frost or dew, cubic ice )

aerosol effects

curvature effects

slow drift by changing the ice shape/amount

slow drift of detector signal

13

Example sounding of contamination at cloudy condition

A thick cloud causes the contamination. After passing the cloud, the oscillation is caused because the

contaminated air flow into the mirror.

It seems that descent data also indicate higher frost point than an expected value.

RHice=100%

Photo detector directly

detects cloud particles

contamination

Cloud condition at release time

14

Example sounding of contamination by ethanol

An ethanol to cool the heatsink (hot side of Peltier) can cause the contamination.

It is suspected that the evaporated ethanol breaks frost layer on the mirror. The mechanism of this

behavior is not clear, but this behavior is reproducible in chamber.

We decided that we do not use ethanol for the additional cooling.

ethanolmirror

Mirror temperature is lower

than the expected frost point.

The scattered light signal

become weak.

Mirror temperature is lower

than the expected frost point.

The scattered light signal

become weak.

In chamber

Summary

1. Performance of SKYDEW(2nd product model)

SKYDEW can measure dewpoint/frostpoint up to 30 km at nighttime, and up to 25 km at daytime

(depending on the cooling limitation by Peltier)

2． The product model of SKYDEW is presale in Japan at present.

Only JMA has conducted the sounding with SKYDEW. Because several projects in Japan

are postponed because of COVID-19, many soundings with SKYDEW will be conducted at tropic and

at Antarctica (Syowa station) next year.

3. Action item toward GDP

Data processing, smoothing and time-lag correction (under consideration)

Uncertainty budget (under consideration)

15

For more technical information, contact me, “sugidachit@meisei.co.jp”

If you have interest in purchasing SKYDEW, access the Meisei web site below

https://archive.meisei.co.jp/english/support/contact/

AcknowledgeWe thank JMA (Japan metrological agency) for providing the SKYDEW data at Tateno, 21 October 2020.